cartoon shadows are saturated

RN

How do you pick a shading tone in your colour picker when working on something in a cartoon colouring style?

I'm having trouble with this, so I'm looking for a "formula" or reliable method. Then I found this:

Uniformly coloured surfaces passing between different levels of illumination
under a single light source are represented by image colours that move along a
line of uniform saturation, such that chroma increases as the value increases.

[...]

In digital work we can easily create uniform saturation series by keeping the Hue angle (H) and HSB Saturation (S) constant (in effect keeping the ratio of R/G/B constant), while the HSB brightness (B) decreases. Uniform saturation series are easily created in Photoshop using the colour picker, which allows S and B to be directly manipulated.
In Photoshop shading series can also be created by various shortcut methods such as:

• varying the opacity of a layer in normal mode over a layer of black, or
• varying the opacity of a layer of black in multiply mode over a layer
• by placing a coloured layer in multiply mode over a greyscale image.

From: http://www.huevaluechroma.com/101.php#shadingseries

So that's a reliable and scientifically proven method, but the dark tone that you find is less saturated \ less vibrant than the light tone that you're using. It takes away some of the appeal.

I colour-sampled a few reference pictures and found that people rely on saturated shadows when working on stuff of this style (that is, in the HSB picker the saturation value of the shadow colour is equal or greater than that of the light colour):


From "Youkai Watch" (EDIT 28-01-2016) This image uses Multiply for the shadows, check this post below.


From http://danikruse.tumblr.com/image/114147243522

Using the L-a-b picker in Photoshop, if I read what's the L value (the perceived lightness) of the dark tone obtained from that scientific method above, then pick the light tone again and drop the L slider of the light tone until its L value is the same as that dark, the new dark tone found this way is slightly more saturated.



The light tone is the same, but the dark tones are picked with different methods.
What I get from this is, if you have a "centre light" tone and you want to find a darker saturated shadow tone for it, changing the L slider in the L-a-b picker seems to give something reasonable.
I wonder if there's some theory to cartoon shadowing that could help with this.

Muzzoid

If you want saturated shadows, and to do it correct, you want to have a saturated ambient colour. You can approximate this by having a shadow layer set to multiply and tinting it the colour of your ambient, though to do it 100% correct would require a more complicated layer stack.
LAB mode is a theory space that is designed to maximize the possible colour gamut in an image format, and has inherent benefits in image processing, but as a theory doesn't explain jack shit about light. Every-thing you need to understand can be understood with simple rgb colour math .

RN

Hey Muzz, thank you so much. The braincells started working. Getting the terminology right seems to help with this.
The "problem" that looks like the desaturation of shadows is described in that same quote from David's website.

There's a TL,DR at the end of this post.
There are two measurements of the "purity" of colour involved here: saturation and chroma. It seems easier to understand these when thinking about the RGB and HSB sliders that you find in art programs.

- Saturation (in HSB) is measured by how spread apart the RGB channels are in relation to the slider range delimited by the HSB brightness value.
- The chroma of a colour (the modern meaning of it) can be thought of how spread apart its RGB channels are in relation to a grey of the same lightness. In practice, this means that colours with more different RGB channel values have stronger chroma than colours with channel values that are closer together. This comparison happens in the full slider range, the HSB brightness doesn't influence here.

With David's methods for finding shadow tones, when you lower the HSB brightness of your colour, this brightness acts as a factor that scales the channel values and preserves their ratio (how strong they are in relation to each other) within that range delimited by the brightness value. The slider name below is "V", but it's the same as the B in Photoshop.

The "problem" is this: as you drop the brightness, when you look at the full slider range, notice how the channel values are becoming closer to each other. Like in that quote, this means that the chroma of the colour is decreasing. The colour will appear to have less "vibrance".
The saturation is preserved (their ratio within that brightness range) and the chroma changes.

When using the L* slider in L*a*b* mode, the chroma is preserved and the saturation changes.

Notice that when sliding down the L* value all channels are 'offset' by equal amounts. When darkening a colour, this has the effect of increasing its HSB saturation.
After colour-sampling some more cartoons, it seems that while "real world" shadows preserve saturation, cartoon shadows preserve chroma (look how shadow tones are more saturated):

I noticed the same when sampling some value series in the Munsell system. The practical effect of all of these is that all the RGB values go down by similar, if not equal amounts in some shadow tones.

TL, DR

To pick a shadow tone in a "light \ shadow" tone pair for a cartoon-style rendering, pick your light colour then do either of the following:
- In a L*a*b* colour picker, lower the L* value of that light colour. The resulting tone is your shadow. Or...
- Lower all your RGB channels by some fixed amount. Pick a number like "40", for example, and subtract it from all channels. The resulting tone is your shadow.

The L* value seems to be useful for a lot of colour value stuff. It approximates (so it's not perfect) the way that humans perceive the luminance of colours. I'll study some more to see how it can help.

meguskus

It doesn't matter much whether it's realistic or stylized, shadows are always more saturated than most people think they are. What gives them colour is the reflected light from the sky or the object that is casting the shadows, often a mix of both. That is why shadows are often a bluish hue if the sky is clear and blue and if the lightsource is warm (indoors incandescent lighting for example), the shadows are generally warmer in hue. Never are the shadow parts solely a darker value than the light parts of the same object, because the environment as well as the object's material itself influence the shadow a lot.
I suggest you read Gurney's Color and Light for some information.

RN

meguskus said:

Never are the shadow parts solely a darker value than the light parts of the same object, because the environment as well as the object's material itself influence the shadow a lot.

That's absolutely true. In some cartoon renderings I don't see that being used, though.

Just dropping the L* slider value to get a shadow tone will only work in some cases. You guys mention how important coloured illumination (or coloured ambient light) is. After analysing some more toon media and those previous images again, it's plausible that layers in multiply mode were used to get a coloured ambient light effect.

The strongest example is in that image from "Youkai Watch", every character is shaded by the same colour in multiply mode.
The image is this: http://s30.postimg.org/ul74rqkq9/yokai_watch.jpg
The shadow colour is this: bf9fa4 (hex)
If you paste that image into Photoshop, make a new layer, set it to Multiply mode and paint with that colour with a hard brush, all the shadow tones in those characters will match.

EDIT: Another example of a full scene lit by the same colour in multiply mode (except for the two foreground characters).
The image is this:http://s30.postimg.org/jz5e99ve9/svfoe.png
The shadow colour is this: a9d1e2

In the other images there isn't that same connection. Each light \ shadow pair seems to use a different multiplying colour, I can't see a pattern there other than "to make it look good".

A couple of ways to use a 'multiply layer' that look interesting:

Multiplying a colour on top of itself

- Paint a shape in a layer using a certain colour.
- Create another layer on top of that one, set it to Multiply mode and with Clipping Mask on.
- Use a brush also in Multiply mode and paint on that top layer with that same colour.

It's like shadowing the colour and increasing its "flavour" at the same time. Seems to give the effect of transmitted light (or SSS), like what happens with glaciers and foliage.

Multiply shadowing

(EDIT: This is what Muzz was talking about, I think.)

- In a layer, paint a shape in a chosen light colour.
- Create a layer on top of that, set it to Multiply mode and turn on Clipping Mask for it.
- Pick white (255,255,255), then using the pop-up picker in Photoshop, lower the HSB brightness of that white thinking of it as a percentage. 85% of brightness means that the light colour will be darkened by 85%.
Pick a hue, then pick a saturation. While changing the saturation the apparent lightness of the colour also changes. You can look at the L value to have some bearings on how dark the tone is becoming. The darker the tone is in comparison to white, the darker the shadow will be.
All of this is just so there's less of a surprise about the resulting colour when the blending happens. You have a feeling of how much it will darken and in what direction (i.e hue) it will take the shadows.

Tone pair and smooth tests, aiming for a stylised look.
It's not a penis.


To make corrections in the shadow layer you can just erase parts of it. But if you're using several shadow tones then you need to 'lock transparency' and paint over with the shadow colour that you started with. It's difficult to keep track of all these shadow tones, so using swatches or a scratch document as explained here seems to help.

Even though it's being used for "cartoon" style shadows, the multiply mode forces you to think more like light behaves. More on that here: http://www.huevaluechroma.com/104.php

This is also interesting, guesstimating the effects of coloured lights: http://kbladin.se/tools/color_changes.php

Wolthera

I am 90% sure that some of those earlier examples just use HSL sliders instead of HSB.

Also, check pixel art dedicated websites, they have some basic examples of stylised ramp building. In particular, a common advice is to hue-shift towards blue because typical stylised colors are made with the assumption that shadows are blue(and some even have it that lights are yellow). This is based on 18th century color theory and should be seen as an aesthetic method rather than an attempt at mimesis.(Mimesis, btw, is my new favourite word)

I made a list of stylised ramping methods a few years back under the "Basic ramping formations and Stylised colours:" header here: http://theratutorial.tumblr.com/post/68488882915/colour-theory-more-muddy-than-colourful

RN

I found a couple more resources. These are in Japanese, but you can get the gist by translating with Google etc.

# 1

From: https://howto.clip-studio.com/library/page/view/illuststudio_making_c_028_005

Artist uses a layer in multiply mode for shadowing. The same shadow colour is used for the whole scene.
The base colours represent the scene fully lit (thought they do add some lighter details later), then the shadow layer works on top of that.
Bounced light is produced by erasing parts of the shadow layer so that it darkens less of the base colours. They also stack multiple shadow layers (all in multiply) to emphasize the depth in parts like cloth wrinkles and creases.


# 2

From: http://east99.nobody.jp/animenuri02.html

This artist also uses layers in multiply mode for shadowing, but for each different material he uses a different multiplying colour.